Defuddle
Extract clean, distraction-free Markdown from noisy web pages.
Steph Ango (Kepano) is credited as the original ai skill creator. Setup notes, fit assessments, and cautions below are Obsidian Mate editorial guidance. How we review resources
Overview
A web extraction utility skill by Steph Ango (Kepano) that strips web page clutter into clean, distraction-free Markdown. Use it when an agent needs to extract the main readable content of a web page before saving a Markdown note. This page treats the skill as an instruction package for an existing agent host, not as an Obsidian plugin or a grant of new permissions.
Key Features
- Noisy web element stripping
- Clean heading and image preservation
- Obsidian Clipper compatibility
How to Install Defuddle
- 1Open the source repository and read its current installation instructions for your agent host.
- 2Install the complete skill folder, including any referenced files, into the host's documented skills location.
- 3Start a fresh agent session and confirm the skill is discoverable before granting vault write access.
- 4Run the first workflow in a disposable vault and review the resulting file diff.
How to Use Defuddle
Test one public article, compare the extracted title and body with the page, remove navigation residue, and preserve the source URL and access date in the resulting note.
When to Use This Skill
Use it when an agent needs to extract the main readable content of a web page before saving a Markdown note.
End-to-End Check
Test one public article, compare the extracted title and body with the page, remove navigation residue, and preserve the source URL and access date in the resulting note.
Compatibility and Safety
- Extraction can fail on paywalled, script-heavy, or unusual pages. The cleaned text is not proof that every figure, caption, or attribution was retained.
- The host agent still controls tool access, filesystem scope, and confirmation behavior.
What does this skill do?
Cleans HTML web page content into distraction-free Markdown before saving to an Obsidian vault. Use it when an agent needs to extract the main readable content of a web page before saving a Markdown note. The agent host remains responsible for file access, tool availability, and confirmation before writes.
Best for
- Clipping articles into Obsidian without clutter
- Converting blog posts and documentation to clean Markdown
- Removing ads, popups, and sidebars
How it works
- Fetches web page HTML
- Removes header, footer, sidebar, and script elements
- Outputs clean article body in Obsidian Markdown
License & Attribution
COMMUNITY RESOURCE Author: Steph Ango (Kepano) License: MIT Repository: https://github.com/kepano/obsidian-skills
Details
- Author
- Steph Ango (Kepano)
- License
- MIT
- Type
- Agent Skill
- Compatible with
- Claude Code, Codex, ChatGPT, Gemini, Agent Skills
Frequently Asked Questions
Does it preserve main article images?
Yes, core article body images and code blocks are preserved.
Sources and Documentation
This guide was checked against the creator's repository and primary documentation.
Related Obsidian AI Skills
Obsidian Markdown
Teaches AI agents how to use Obsidian-flavored Markdown syntax correctly.
Obsidian Bases
Allows AI agents to manage Obsidian Bases (.base files), formulas, and database views.
JSON Canvas
Allows AI agents to create and edit visual JSON Canvas (.canvas) diagrams.
Obsidian CLI
Teaches AI agents how to interact with Obsidian vaults via command-line interface.
Explore more Obsidian resources
Discover more plugins, themes, templates, and starter vaults to build your perfect Obsidian setup.